hpcups and hpijs driver use different rasterization

[Rüdiger Kupper]

HPLIP provides two drivers for the Laserjet 1515n: hpcups and hpijs:
- hp-color_laserjet_cp1515n-hpijs-pcl3.ppd
- hp-color_laserjet_cp1515n-pcl3.ppd

The new hpcups shall replace the old hpijs, however, it produces very different rasterization results, resulting in a radically different visual impression. Differences are especially prominent for greyscale prints.

Hpcups produces "amplitude modulation" halftones (dots of different sizes at fixed locations). When resolution is set to "Normal", the size of these dots is so large that contours cannot be reproduced with reasonable detail. Heavy aliasing can be easily seen in the circle contours on the Ubuntu Printer test page. Printing photographs in greyscale mode yields badly rasterized images like those known from low-quality newspaper prints (dot pattern is so prominent it dominates image contours). Artifacts are reduced, but still visible at resolution "Best".

In contrast, hpijs produces "frequency modulation" halftones with stochastic dithering (small-sized dots placed at varying densities). At resolution "Normal", it produces evenly colored areas with reasonably sharp edges. Circle contours on the Ubuntu Printer test page appear continuous. Photographs are reproduced at a a much higher fidelity as with the hpcups driver.

Questions:
1. Why do the two drivers use different rasterization?
2. Why is the result of rasterization so bad in the hpcups driver, set at resolution "Normal, Greyscale"?

Note:
Both drivers produce acceptable color prints with much fewer artifacts visible, even at resolution "Normal".

ProblemType: Bug
DistroRelease: Ubuntu 12.04
Package: hplip 3.12.2-1ubuntu3.1
ProcVersionSignature: Ubuntu 3.2.0-29.46-generic 3.2.24
Uname: Linux 3.2.0-29-generic i686
ApportVersion: 2.0.1-0ubuntu12
Architecture: i386
CurrentDmesg: [ 63.368304] init: plymouth-stop pre-start process (1700) terminated with status 1
Date: Mon Aug 13 16:38:16 2012
Lpstat: device for Farblaser: hp:/net/HP_Color_LaserJet_CP1515n?zc=NPIE56824
MachineType: exone D156200
Papersize: letter
PpdFiles: Farblaser: HP Color LaserJet cp1515n pcl3, hpcups 3.12.2
ProcKernelCmdLine: BOOT_IMAGE=/boot/vmlinuz-3.2.0-29-generic root=/dev/mapper/lvmvg-Root ro quiet splash vt.handoff=7
SourcePackage: hplip
UpgradeStatus: Upgraded to precise on 2012-04-26 (108 days ago)
dmi.bios.date: 04/23/2004
dmi.bios.vendor: FUJITSU SIEMENS // Phoenix Technologies Ltd.
dmi.bios.version: 4.06 Rev. 1.09.1562
dmi.board.name: D1562
dmi.board.vendor: FUJITSU SIEMENS
dmi.board.version: S26361-D1562
dmi.chassis.type: 6
dmi.chassis.vendor: exone
dmi.modalias: dmi:bvnFUJITSUSIEMENS//PhoenixTechnologiesLtd.:bvr4.06Rev.1.09.1562:bd04/23/2004:svnexone:pnD156200:pvr:rvnFUJITSUSIEMENS:rnD1562:rvrS26361-D1562:cvnexone:ct6:cvr:
dmi.product.name: D156200
dmi.sys.vendor: exone

[Till Kamppeter (till-kamppeter)]

This is a problem of the upstream design of the HPLIP printer drivers by HP, not of the Ubuntu packaging of the drivers. Therefore I am adding an upstream task to make your bug report visible to the developers at HP.

Can someone of the HPLIP developers at HP look into this and tell why the two driver flavors (hpcups and hpijs) use so different halftoning algorithms?

[Launchpad Janitor (janitor)]

Status changed to 'Confirmed' because the bug affects multiple users.

[Hasenkopf, Andreas (andi-hasenkopf)]

I'd like to add that printing an image (colored, but mostly black) is fine when using the "normal" mode, but the contrast on the paper is very poor when printing in the "normal grayscale" mode. Black lines are printed as very light gray lines.
I have encountered this on different HP (color) laserjet printers.

The problem affects Ubuntu 12.04 (64bit) and Debian squeeze (64bit).

[gf (gf-interlinks-deactivatedaccount)]

Hello Rudiger,
Thank you for submitting this bug and reporting a problem with the hplip package. You made this bug report in 2012 and Ubuntu and hplip have been updated since then.

Could you confirm that this is no longer a problem and that we can close the ticket?
If it is still a problem, are you still interested in finding a solution to this bug?
If you are, could you let us know and, in the current version, could you run the following (only once):
apport-collect 1036236
and upload the updated logs and and any other logs that are relevant for this particular issue.

Thank you again for helping make Ubuntu better.
G
[Ubuntu Bug Squad volunteer triager]

[J T (elmer.fudd)]

ProblemType: Bug
ApportVersion: 2.20.10-0ubuntu27
Architecture: amd64
CupsErrorLog:
 W [20/Jun/2019:09:16:39 +0100] CreateProfile failed: org.freedesktop.ColorManager.AlreadyExists:profile id \'serverLaser.hpijs-Gray..\' already exists
 W [20/Jun/2019:09:16:39 +0100] CreateProfile failed: org.freedesktop.ColorManager.AlreadyExists:profile id \'serverLaser.hpijs-RGB..\' already exists
 W [20/Jun/2019:09:17:31 +0100] CreateProfile failed: org.freedesktop.ColorManager.AlreadyExists:profile id \'serverLaser.hpcups-Gray..\' already exists
 W [20/Jun/2019:09:17:31 +0100] CreateProfile failed: org.freedesktop.ColorManager.AlreadyExists:profile id \'serverLaser.hpcups-RGB..\' already exists
CurrentDesktop: ubuntu:GNOME
DistroRelease: Ubuntu 19.04
Lpstat:
 device for serverLaser.hpcups: parallel:/dev/usb/lp0
 device for serverLaser.hpijs: parallel:/dev/usb/lp0
MachineType: Hewlett-Packard HP EliteBook 8560p
Package: hplip 3.19.1+dfsg0-1
PackageArchitecture: amd64
Papersize: letter
PpdFiles:
 Error: command ['fgrep', '-H', '*NickName', '/etc/cups/ppd/serverLaser.hpcups.ppd', '/etc/cups/ppd/serverLaser.hpijs.ppd'] failed with exit code 2: grep: /etc/cups/ppd/serverLaser.hpcups.ppd: Permission denied
 grep: /etc/cups/ppd/serverLaser.hpijs.ppd: Permission denied
ProcKernelCmdLine: BOOT_IMAGE=/boot/vmlinuz-5.0.0-17-generic root=UUID=e119eace-ded7-4a54-b62d-d08677374c7c ro splash quiet vt.handoff=1
ProcVersionSignature: Ubuntu 5.0.0-17.18-generic 5.0.8
Tags: disco
Uname: Linux 5.0.0-17-generic x86_64
UpgradeStatus: No upgrade log present (probably fresh install)
UserGroups: adm cdrom dip lpadmin plugdev sambashare sudo
_MarkForUpload: True
dmi.bios.date: 02/13/2018
dmi.bios.vendor: Hewlett-Packard
dmi.bios.version: 68SCF Ver. F.67
dmi.board.name: 1618
dmi.board.vendor: Hewlett-Packard
dmi.board.version: KBC Version 97.4E
dmi.chassis.type: 10
dmi.chassis.vendor: Hewlett-Packard
dmi.modalias: dmi:bvnHewlett-Packard:bvr68SCFVer.F.67:bd02/13/2018:svnHewlett-Packard:pnHPEliteBook8560p:pvrA0001D02:rvnHewlett-Packard:rn1618:rvrKBCVersion97.4E:cvnHewlett-Packard:ct10:cvr:
dmi.product.family: 103C_5336AN
dmi.product.name: HP EliteBook 8560p
dmi.product.sku: WX787AV
dmi.product.version: A0001D02
dmi.sys.vendor: Hewlett-Packard

[J T (elmer.fudd)]

apport information

[J T (elmer.fudd)]

apport information

[J T (elmer.fudd)]

apport information

[J T (elmer.fudd)]

apport information

[J T (elmer.fudd)]

apport information

[J T (elmer.fudd)]

apport information

[J T (elmer.fudd)]

apport information

[J T (elmer.fudd)]

apport information

[J T (elmer.fudd)]

apport information

[J T (elmer.fudd)]

apport information

[J T (elmer.fudd)]

apport information

[J T (elmer.fudd)]

apport information

[J T (elmer.fudd)]

Hello gf.

I am not Rüdiger, the original submitter of this bug, but I am affected by it. Since Rüdiger has not posted to this bug report in over 7 years, I am posting to it to let you know that:

1) It *is* still a problem,
2) I *am* still interested in finding a solution to this bug,
and
3) I would like you to *not* close the ticket.

Is there a better place for me to report this bug? (HP's FAQ at https://developers.hp.com/hp-linux-imaging-and-printing/faqs suggests reporting it here, but this bug doesn't seem to have had any work done on it since Suma Byrappa (suma-byrappa) became an assignee in December 2013).

Thanks.

[Till Kamppeter (till-kamppeter)]

J T, you are posting your auto-generated bug data here but do not post a comment telling what is exactly your observation.

Note that in HPLIP upstream HPIJS is still contained but not further developed. We also do not explicitly support it any more in Ubuntu, we encourage users to switch to hpcups and do even automatic migration.

If you observe any output quality advantage in HPIJS, please report a new bug to HPLIP upstream (https://bugs.launchpad.net/hplip/), so that they can compare and improve the hpcups driver.

Nothing of Ubuntu and its packaging is involved here. So I am closing this bug.

[Rüdiger Kupper (ruediger.kupper)]

Hello J T,
thanks for reporting that this is still a problem. I had originally filed this report in 2012, but since it never moved forward I just resigned and kept using the HPIJS driver, which produces much better (in terms of visually accurate) output. However, Till Kamppeter says that Ubuntu is actively phasing out the HPIJS driver. This means that the HPCUPS driver needs to be improved/fixed.

This bug is (upstream) assigned to Suma Byrappa, so he/she should be the person to contact.

@Suma Byrappa:
Can you confirm that hpcups uses different halftoning algorithms (amplitude vs. frequency modulation). Why is this? Why does it mainly concern greyscale prints, not color? How about adding an option of selecting the algorithm in the officially maintained driver (which is hpcups, as I understand)?

Thanks for your support!

[Rüdiger Kupper (ruediger.kupper)]

For the record: Upstream bug is at https://bugs.launchpad.net/hplip/+bug/1036236. But projects are linked.

[J T (elmer.fudd)]

# Hello J T,

Hi Rüdiger,

# I just resigned and kept using the HPIJS driver, which produces much better
# (in terms of visually accurate) output.

Exactly the same situation here.

# However, Till Kamppeter says that Ubuntu is actively phasing out the HPIJS
# driver. This means that the HPCUPS driver needs to be improved/fixed.

I have the same opinion too. It is precisely *because* I read that hpijs
was being "deprecated" that made me write in and post a comment to your bug
report. (The hpijs driver produces good prints with my HP1100 - the hpcups
driver, on the other hand, produces images that are completely unacceptable).

I have spent an hour reading random HPLIP bug reports and their "resolutions"
and as a result, I unfortunately do not hold out much hope that this bug will
ever be resolved. While I would love to see the hpcups driver "fixed", I'll
continue to use the hpijs driver in the meantime, and if nothing changes
before the hpijs driver is "removed", I'll have to go and buy a non-HP
printer.

Fingers crossed...

[Matt Whitlock (whitslack)]

While debugging an unrelated issue with printing to my HP LaserJet 1012, I switched from the "hpcups" driver to the "hpijs" driver, and I am **BLOWN AWAY** by the quality of its dithering. After seeing the difference, I am absolutely appalled at how poorly the "hpcups" driver dithers grayscales. I had no idea that my cheap little laser printer could produce such great looking graphics until I discovered the "hpijs" driver. The "hpcups" driver is garbage by comparison.

[J T (elmer.fudd)]

On 02/07/2019, Matt Whitlock <email address hidden> wrote:
> While debugging an unrelated issue with printing to my HP LaserJet 1012,
> I switched from the "hpcups" driver to the "hpijs" driver, and I am
> **BLOWN AWAY** by the quality of its dithering. After seeing the
> difference, I am absolutely appalled at how poorly the "hpcups" driver
> dithers grayscales. I had no idea that my cheap little laser printer
> could produce such great looking graphics until I discovered the "hpijs"
> driver. The "hpcups" driver is garbage by comparison.

I totally agree. I love the enforced "upgrade" to something which is
completely unusable. Clearly the people at HP are not using HP
printers, or this would have been fixed (six) years ago.

[Rüdiger Kupper (ruediger.kupper)]

Is the hpcups driver actually being maintained by HP? Or is this a community driver?

[J T (elmer.fudd)]

On 02/07/2019, Rüdiger Kupper <email address hidden> wrote:
> Is the hpcups driver actually being maintained by HP? Or is this a
> community driver?

I don't know (and I don't even know how to find out!) I also don't
know whether the following info has any relevance to your question
or not, but:
https://github.com/twaugh/hplip/blob/master/check-plugin.py
says:
# (c) Copyright 2011-2015 HP Development Company, L.P.
# Author: Suma Byrappa, Amarnath Chitumalla

<confused>

[Rüdiger Kupper (ruediger.kupper)]

Well, at least Suma Byrappa is the person that has this bug assigned, and Amarnath Chitumalla is listes as "may be notified" on this page. The two should probably know the answer.